Guray Ozen fa13c3eea7 [mlir][nvgpu] Fix transposeB in nvgpu.warpgroup.mma (#79271) ...

The #76150 fixed meaning of `transposeB` in NVVM dialect which was
initially implemented with opposite meaning.

This PR fixes the lowering of `nvgpu.warpgroup.mma` to NVVM dialect.

This will fix two integration tests:
gemm_f32_f16_f16_128x128x128.mlir
gemm_pred_f32_f16_f16_128x128x128.mlir

2024-01-25 09:25:43 +01:00

..

CMakeLists.txt

[MLIR][NVGPU] Introduce nvgpu.wargroup.mma.store Op for Hopper GPUs (#65441)

2023-10-05 10:54:13 +02:00

NVGPUToNVVM.cpp

[mlir][nvgpu] Fix transposeB in nvgpu.warpgroup.mma (#79271)

2024-01-25 09:25:43 +01:00